[mrtg] CPU and RAM Utiliziation for Windows Machine

Matt Baer matt at baerconsult.com
Sat Sep 26 04:27:39 CEST 2009


Using the generic template as suggested, I managed to get CPU going on a quad core. Still trying to find a better way to display it so I can see each individual core on one graph, but that's out there on the mailing list now as well. Thank you very much, I think that template may solve the issue I've been having trying to get load values populating in a graph. 

----- Original Message ----- 
From: "Steve Shipway" <s.shipway at auckland.ac.nz> 
To: "Matt Baer" <matt at baerconsult.com>, "mrtg" <mrtg at lists.oetiker.ch> 
Sent: Friday, September 25, 2009 7:58:43 PM GMT -06:00 US/Canada Central 
Subject: RE: [mrtg] CPU and RAM Utiliziation for Windows Machine 


I'd really suggest using cfgmaker templates to generate the various Targets. This way you will get all-numerical OIDs (so no problems with unrecognised short symbolic names) and also the host cfgmaker template in the website probes for th existance of the various OIDs before creating the Target definition, so you wont get one that doesnt exist. 

The cfgmaker documentation tells you the details, but basically you just use 
cfgmaker --host-template=template.htp community at hostname 
if your monitored host is 'hostname' with snmp community 'community' and the host template is called 'template.htp'. An interface template is similar, except that it is run once PER DETECTED INTERFACE rather than just once. 

Note that there are some problems with using SNMP - 
1. You need to make sure it is enabled 
2. You need to make sure the apporpriate sections are enabled - eg with Linux you need to have the appropriate lines in the /etc/snmp/snmpd.conf to enable disk space checks, load average checks, and so on 
3. Different OSs support different parts of the OID tree. For example, Windows and Linux give different places for data, and so do Solaris, Cisco and so on. Plus different vendors have vendor-specific extensions which may or may not be useful. 
4. You need to set up your permissions correctly to allow the monitoring server rights to view the OID tree correctly 
5. With Windows in particular, the Storage mappings can be re-enumerated on the fly if you plug in or remove USB devices. This can mess up storage graphs done via SNMP. You may be better off with something the like mrtg-storage check plugin which uses SNMP but confirms the table has not been re-enumerated. 

If you run with the all-hosts host template from www.steveshipway.org/cfgmaker and it doesnt generate any Targets then check your SNMP configuration. 

Steve 


From: mrtg-bounces at lists.oetiker.ch [mrtg-bounces at lists.oetiker.ch] On Behalf Of Matt Baer [matt at baerconsult.com] 
Sent: Saturday, 26 September 2009 9:58 a.m. 
To: mrtg 
Subject: [mrtg] CPU and RAM Utiliziation for Windows Machine 




I'm having some issues with monitoring the CPU and RAM of a Phenom machine running Vista. I'd like to watch all 4 cores. I'm sure it's a problem with OID. But don't know what to switch to. This is what I get in my logs: 

SNMP Error: 
Received SNMP response with error code 
error status: noSuchName 
index 1 (OID: 1.3.6.1.4.1.2021.4.6.0) 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/mrtg/attachments/20090925/aa9b5472/attachment.htm 


More information about the mrtg mailing list